Transformers

Les Transformers ont changé la donne pour l’IA générative, en permettant aux modèles de génération de texte de mieux comprendre la langue.

Vous disposez d’une application interactive qui visualise le mécanisme d’auto-attention. La phrase est : Transformers excel at grasping the context of text, which allows them to generate more coherent responses., où le mot "them" renvoie à un autre mot de la phrase.

Cliquez sur le mot "them" dans l’application pour explorer à quel point le modèle le relie à chaque autre mot.

Cet exercice fait partie du cours

<cours>Concepts d'IA générative</cours>

Voir le cours

Exercice interactif pratique

Transformez la théorie en action avec l’un de nos exercices interactifs

Commencer l’exercice

Cet exercice fait partie du cours

<cours>Concepts d'IA générative</cours>

DébutantNiveau de compétence

4.8+

7684 reviews

Commencer le cours gratuitement

Familiarize yourself with the concept of generative AI and its ability to create content is introduced, along with its real-world applications and limitations. You'll delve into the differences between traditional machine learning models, generative AI, and artificial general intelligence (AGI), and explore the key factors driving the development of generative AI.

Exercise 1: Qu'est-ce que l'IA générative ?Exercise 2: Découvrir l'IA générative Exercise 3: Travailler avec un modèle d'IA générative Exercise 4: Applications de l'IA générative Exercise 5: L'IA générative dans le paysage du machine learning Exercise 6: Choisir le bon outil pour la tâche Exercise 7: Différencier l'IA générative du ML traditionnel Exercise 8: Plus de chiens et de GAN Exercise 9: L'évolution de l'IA générative Exercise 10: Faire avancer le domaine Exercise 11: Transformers

Exercice actuel

Exercise 12: Percées de l'IA générative

In this chapter, we cover the essential steps in creating generative AI models: research and design, data collection, model training, and evaluation. We examine the significance of diverse datasets and advanced training techniques, as well as various evaluation methods, while discussing their strengths and limitations.

Exercise 1: Model design and data collection Exercise 2: Research and design Exercise 3: A dataset windfall?Exercise 4: Model training Exercise 5: Factors impacting model training Exercise 6: Techniques for developing specialized models Exercise 7: Model evaluation Exercise 8: PixAIr animation Exercise 9: Helping Intellibrand evaluate

This chapter focuses on the responsible use of generative AI. We discuss the challenges and strategies to mitigate social bias, intellectual property and privacy issues, and ethical considerations to prevent misuse. We conclude by exploring the immense potential and risks of Artificial Generative Intelligence (AGI), along with the approaches to control its outcomes.

Exercise 1: Evaluating and mitigating social bias Exercise 2: Sources of bias Exercise 3: Detecting and mitigating bias Exercise 4: Copyright and ownership Exercise 5: Legal and privacy considerations Exercise 6: Identifying IP ownership claims Exercise 7: Responsible generative AI applications Exercise 8: Usage principles Exercise 9: Implementing responsible AI practices Exercise 10: Artificial general intelligence (AGI)Exercise 11: AGI characteristics Exercise 12: AGI outcomes

Chapter 4 examines the potential, impact, and integration of generative AI into human workflows. It discusses key contributors to AI development, from universities to companies, and explores societal adaptations to AI. It delves into AI's implications for productivity, job dynamics, education, media, entertainment, scientific advancements, and ethical considerations.

Exercise 1: Bringing new AI into old workflows Exercise 2: Leveraging generative AI for data science Exercise 3: Designing a new marketing campaign Exercise 4: Progress in generative AI Exercise 5: To open source or not to open source Exercise 6: Super Duper GANs Exercise 7: Preparing for a future of generative AI Exercise 8: Generative AI in education Exercise 9: The value of skills in an AI world Exercise 10: You made it!